[Raijin Thunderkeg]

Hi, All.

I'm modelling the distribution of dye in the water of a water-gas system. The volume fraction distribution of this 2-phase system is as shown below. I insert
a water-specific additional variable and a transport equation is set for it and a kinematic diffusivity has been set. After 20 s, the distribution of dye is shown
as below. It can be seen from the contour that the dye spreads into the gas phase. This is repugnant to common sense because this additional variable can
exist only in water phase in physics.

My question is : What is the reason for this ? Interphase transfer ? How to solve it?

[ghorrocks]

Unless you have set the diffusivity to zero in the gas phase it will diffuse into the gas as well. And free surface models are not perfect - they have a small amount of numerical diffusion at the free surface anyway so a little bit of your dye is going to end up in the gas from that.

And does it matter anyway? Providing it does not artificially spread the stuff to the liquid phase then you can easily ignore it in post processing anyway.

So (in increasing degree of effort):
* Check to see whether it matters.
* make sure you have set the gas diffusivity to zero.
* Do a mesh refinement study to see if it improves.
* Have a play with the free surface modelling parameters and see if you can reduce the diffusion across the interface.

[Raijin Thunderkeg]

Thanks for your reply. It has troubled me for a long time. It seems I misunderstand some concept about multiphase. The phasic transport equation for additional variable is not solved in phase-specific domain.